Semi-autonomous driving

Designing control systems for autonomous vehicles in mixed-autonomy situations


Project Members: Ahmed Allibhoy


Traffic jams are responsible for billions of dollars in losses and wasted energy. Although traffic jams occur when vehicle density exceeds a certain threshold, they may also occur without any trigger at all. The instability of equilibrium flow in traffic systems is a fundamental consequence of the way humans drive -- their reactions to drivers in front of them causes small perturbations to propagate down a single chain of vehicles and amplify.

Autonomous vehicles present an opportunity to mitigate this problem as they can be programmed to behave in a manner as to prevent disturbances from equilibrium flow from magnifying, thus halting the propagation of a traffic wave. However, it will be a long time before autonomous vehicles fully replace human drivers, and in the near future human drivers and autonomous vehicles will be shar ing the same roads.

Research suggests that in mixed autonomy situations, it is still possible for autonomous vehicles to improve traffic flow . If we can program autonomous vehicles to influence people to drive better, they may attenuate traffic waves and improve vehicle throughput. Advances in autonomous vehicle technology provide us with the opportunity improve safety, get people to their destinations faster, and dramatically increase the capacity of existing highways without requiring major overhauls or additions to our transportation infrastructure.